logginghas substanhally reducedthe time reqmredfor theseareasto return to their approximatepreharvestbulk densities. LITERATURE [] CITED DiCr,•ON, D.P. 1976.Softcompaction aftertreelen•h skiddingin northernMississippi.Soil So. Soc. Am. J. 40:965-966. DRYNESS, C.T. 1965. Soft surfacecondition follow- ingtractorand high-leadloggingin the Oregon Cascades.J. For. 63:272-275. FroTH,J., ANt) G. M•H¾. 1989. Skid trails and •heir effect on growth and managementof youngPinusradiata.N. Z. J. For. Sci.19(1):2228 FaOEHtICH,H.A., D.W.R. MILES, ANt) R.W. ROE- ruNS.1985.Soilbulk densityrecoveryon compactedskidtrailsin centralIdaho.SoilSd. Soc. Am. J. 49:1015-1017. HAMMOND, S C 1989 Impactsof timberharvestingon softsandresidualtreesm Indiana'sMorgan-MonroeStateForest.M.S. thesis,Purdue Univ., Dep. For. and Nat. Resour.93 p. HATCHELL,G.E., C.W. R•LSTON,ANDR.R. FOIL. 1970.Soil disturbances in logging--Effects on soilcharacteristics and growthof loblollypine in the Atlantic CoastalPlain. J. For. 68(12):772775. HOMOYA,M.A., D.B. Am•LL, J.R. ALDmCH,AND T.W. POST.1984.The natural regionsof Indiana. Proc. Indiana Acad. Sci. 94:245-269. MACE, A.C., JR. 1971. Recoveryof forest soils from compactionby rubber-tired skidders. Minn. For. Res. Notes No. 226. Univ. Minn., St. Paul. REISINGER,T.W., G.L. SIMMONS, AND P.E. POPE. 1988.The impact of timber harvestingon soil propertiesand seedlinggrowthin the south. South.J. Appl. For. 12(1):58-67. REISINGER• T.W., ANDP.E. POPE.1991.Impactof timberharvestingon residualtreesin a central The Use of Tree Shelters with Northern Red Oak Natural Regenerationin Southern hardwood forest in Indiana USDA For Serv Gen Tech Rep NE-14882-91 SIMMONS,G.L. 1987. Effectsof soil compaction and vesicular-arbuscular mycorrhizaeon root dynamicsof yellowpoplarand sweetgumseedlings.Ph.D. diss.,PurdueUniv., West Lafayette, IN. 148p. SOILCONSERVATION SERVICE. 1981.Softsurveyof Morgan and Monroe Counties, Indiana. USDA. 188p. THORVD,D.B., AND S.S. FInSSELL,JR. 1976. Time changesin soil density following compaction under an oak forest. Minn. For. Res. Notes No. 257. Univ. Minn., St. Paul. TwoRosra, T.J., J.A. BVRGER,AND D.W. SMITH. 1983.Soil textureand bulk densityaffectearly growth of white oak seedlings.Tree Plant. Notes, 34(2):22-25. WERT,S., ANDB.R. THOMAS.1981. Effectsof skid roadson diameter,height,and volumegrowth in Douglas-fir.SoilSci.Soc.Am. J. 45:629-632. shelter is sealed with mounded soil (Frearsonand Weiss 1987). Studiesof the operationaluse of shelters have focused on hardwood plantation establishment,with oak speciesreceivingmuch of the attention. The use of shelters has been most New England quicklyadoptedwhere investmentin planting of hardwoodsis already a part of management,as in GreatBritain. A studyof sheltersusedfor proDavid B. Kittredge, Jr. andMatthew J. Kelty, Department of tectingplanted northern red oak in has shownincreasedheight Forestry andWildlifeManagement, UniversityofMassachusetts, Michigan growth over two growing seasons Amherst,MA 01003;andP. Mark S. Ashton, YaleSchool of (Lantagne1990). Red oak is the most commercially Forestryand Environmental Studies,New Haven,CT 06511. importanthardwoodspeciesin southern New England,occurringin mixed ABSTRACT. Theuseof treeshelters with bybrowsing mayhavepotential forensur- hardwoodstandsprimarilywith black naturally established redoakseedlings was ingsuccessful redoakregeneration without birch,red and sugarmaple,and other oak species.Natural regenerationofstudied in southern New Englandmixed- thecosts of planting. ten becomes established in the underNorth.J. Appl.For.9(4):141-145. hardwood stands where deerbrowsing limstory of these stands,but seedlings itedregeneration height growth.Seedlings maybe preventedfrom developingin of threesizeclasses (basaldiameters of 5 Theuse oftree shelters forprotecting height becauseof repeatedbrowsing mm, 8 mm,and 15 mm)werecut offjust by white-taildeer (Kittredgeand Ashabove ground levelduringthedormant sea- planted hardwood seedlings from ton, in prep.). Under these circumsonto induce sprouting. Plastictreeshel- deer browsing has become widestances, savannah-like conditions can terswereplaced overonegroupin eachsize spreadin GreatBritainoverthe past10 develop, in which the understory class; another groupwasleftwithoutshel- years(Potter1991)and is presentlybetersbutwasprotected fromdeerbrowsing ing studied in many other regions growing spaceis usurpedby woody shrubsthat spreadvegetatively,such byfencing.In thefirst growingseason, (Windell 1991).Tree sheltersare tubes as mountain-laurel,or by herbaceous seedling sprouts insideshelters in thetwo made of translucentplastic;they are largersizeclasses haddoubletheheight availablein square,circular,or hexag- fernsand grasses(Horsley1988),further inhibitingoak seedlingdevelopgrowth(withsometerminals growingout onal shapes,generally7 to 15 cm in ment. The potentialexistsin thesesitofthe150-cmtallshelters) butlessdiame- width and 100 to 150 cm in height. uationsfor using shelterswith natuter growth,compared to sproutsoutside They are held in placeover individual rally established oak seedlings to shelters. Theheightdifference wasmain- seedlingsby stakesdriven into the tamed but not increased over the next 2 ground. Tree shelterswere originally preventbrowsedamageandaccelerate developedin GreatBritainin 1979by heightgrowth.It would be possibleto growingseasons.Height growthfor placesheltersdirectlyover the seedsprouts fromthesmaller sizeclass wasnot Graham Tuley (Tuley 1983, 1985), lings, but the stemsare usually de•ncreased byshelters. Theuseof treeshel- principallyto protectseedlingsfrom terswithlargediameter seedlings stunted browsing.They providean additional formedfrom repeatedbrowsing.An advantage by alteringthemicroclimate alternative is to cut the stems back to insidethe sheltersuchthat seedling thegroundduringthe dormantseason overthatof to stimulate the development of a 1The authorsare gratefulto YaleUniver- heightgrowthis increased sity and the Massachusetts Metropolitan unbrowsed, unshelteredseedlings. seedlingsprout within the shelter. Temperatureand humidity levelsare This approachto the use of shelters District Commission(MDC) for use of the maybe morecompatiblewith the lowstudysites,and for the adviceand assis- higher inside shelters(Potter 1991), and carbon dioxide levels are also ininvestmentmanagementpracticedin tanceof BruceSpencer,Paul Lyons, and ThomKyker-Snowman of the MDC. creased,especially if thebottomof the many areas, where planting is not NJAF9(1992) 141 generally done The use of shelters with oak naturalregenerabon has not beenreportedin the literature. The objectiveof this studywas: (1) ment consistedof •nstalhngshelters immediatelyafter clipping.The other group was left unprotectedby shel- to determine the effects of tree shelters browsingby fencing. Shelterswerepurchased froma Britishcompanyandwerenot modifiedin any way fo• use in this study. They were made of light brown, doublewalled corrugatedsheetpolypropylenefoldedintosquaretubes8 cmwide on the heightand diametergrowthof naturallyestablished red oak seedling sproutsof different initial sizes, and (2) to determinethe effectsof shelters on the phenologyand shoot growth patterns of seedlingsand seedling sproutsduring the first growing sea- shelters,and 25 were protectedfrom deer browse by an exclosureconstructed with multiple strands of ters, but was protectedfrom deer barbed wire. A total of 48 individualsin the large size class were measured; 24 were clippedat groundlevelin the dormant season,and 24 were left unclipped In eachgroup of 24, 12 were protected with shelters, and 12 were left without shelters. All trees were within 5-strand electric deer fence. and 150 cm tall. Shelters were held in place by wire fastenersattached to son. METHODS Our studies used natural red oak re- that point, with no data collectedfor the third growingseason.Most com- Measurementsof height and basal diameter(at 3 cm abovethe ground) were madeat the end of eachgrowing seasonfor sprouts developing from clippedseedlingsof all sizeclassesIn all cases,mortalitywas noted.If more than one sprout developed from a stump, the tallestone was measured For sproutsand seedlingsin the large size class,height measurements were madeweeklyduringthe firstgrowing season.At the endof the firstgrowing mercial season, the total number of leaves and 100-cm-tall wooden stakes. Soil was not mounded around the base. Maintenance of shelters was done as neces- generationof three size classes,with mean basal diameters of approximately5 mm, 8 mm, and 15 mm; they sary, consistingmainly of securing wire fastenersthat were loosenedby are referred to hereafter as small, me- many sheltersin the large size class duringthe secondwinterof the study; dium, and large size classes,respectively. Two study siteswere used. In both cases,regenerationin severalrecentlycreatedgaps(0.2-0.5 ha in size) was selectedfor study. The existing a wind. A severe windstorm this treatment blew down had to be abandoned shelter manufacturers at have forest in both cases was mixed hard- now incorporated plastic fasteners the height of each flush of growth wood in composition,with northern red oak, red maple, and black birch predominantin the overstory.Stands were approximately60-80 years old, and even-agedin structure. which are more secure than the earlier were measured. The first location was in northeast- ernConnecticut at theYale-MyersForest, on a relatively good site with a wire designs used in the present study. The sample size of individuals in eachsizecategoryvaried,basedon the numberpresentthat met the size cri- RESULTS Height Growth of Different Size Classes terion. In the small size class, 200 For individuals in the small size seedlingswere clipped,with 100 protectedby treesheltersand 100left unprotectedby shelters,but surrounded by a deer exclosurecomprisedof a class,no significantdifferencewas detected between height growth of very stony fine sandy loam (USDA Soil Conserv.Serv. 1981). Understory vegetationwas dominatedby mountain-laureland seedlingsof red maple and blackbirch. Regeneration at this 5-strand for any of the firstthreegrowingsea- location was in the medium size class, Table1. Height(cm) and basaldiameter(mm) of clippedred oak naturalregenerationin threesizeclasses baseduponinitialbasaldiameter(small= 5 mm, medium = 8 mm, large= 15 mm).a northern red oak site index of 65 (Carmean 1978). Soils were well-drained and moist, and classified as a Paxton with a mean diameter of 7.8 mm [standard error (s.e.) = 0.3 mm]. Counts of growth rings showedthat stemswere 7 to 11 yearsold; rootsmay havebeen older, becausethe stems may have died backand resprouted,as oftenoccurswith oak seedlings. The second location was on the Quabbin Watershed Reservation in central Massachusetts. loam (USDA Soil Conserv. Serv. 1967) that was excessivelydrained. Understory vegetationwas dominatedby lowbush blueberry. Regenerationin both the large and small size classes was used at this site. Mean basal di- ameter of the small categorywas 4.9 mm (s.e. = 0.10 mm), with stemages ranging from 4 to 8 years old, and meanbasaldiameterof the largecategory was 15.1 mm (s.e. = 0.4 mm), with stemagesfrom 8 to 12 years. Seedlingsin each size class were clipped3 to 4 cm aboveground level during the dormant season,and then wererandomlydividedinto two treatment groups. For one group, treat- 142 NJAF9(1992) seedlingsin the medium size class were clipped; 50 were protectedby sons (Table 1). However, the mean basaldiameterof sproutsinsideshel- Small Medium Shelter No shelter Shelter 99 22.3 99 23.4 ns 47 75.1 (1.0) (1.0) 3.3 4.1'* (0.1) (0.1) No shelter Large Shelter No shelter 25 32.9** 12 136.8 12 67 4** (4.6) (2.0) (11.0) (6 0) 3.9 5.0** (0.2) (0.3) 32 17 Year 1 n Height The soil was a Scituateextremely stony fine sandy sproutsinside and outsideof shelters, electric fence. A total of 75 Diameter Year 2 n 94 96 Height 36.3 37.6 ns (2.0) (1.6) Height growth 13.7 (1.4) 14.2 ns (0.9) Diameter 102.5 9 186.6 (6.7) (4.9) (10.3) (7 0) 23.2 (3.7) 16.9 ns (2.9) 31.9 (7.8) 32 8 ns (3 3) 104 0'* 3.6 6.0** 4.2 5.6* 9.9 11 4* (0.1) (0.2) (0.2) (0.5) (0.7) (0.4) 79 84 Height 46.8 51.2 ns 138.6 (2.9) (2.4) (16.1) 14.0 (1.4) 13.7 ns (1.2) Diameter 9 -- 50.5** Year 3 n Height growth -- 10 14.8 (5.1) 14 66.2** (7.1) 15.9 ns (3.7) 3.7 6.6'* 5.4 7.8* (0.1) (0.2) (0.5) (1.0) Resultsof t-testfor differencebetweenmeansin the samesize classare givenas: ns (P > 0.05), * (P < 0.05); ** (P < 0.01). Standarderrorsare listedin parenthesesbeneatheach mean. ters was significantlysmaller than for sproutsoutside sheltersfor all three outsideof sheltersover the 3-year pe- years. In the largesizeclass,heightgrowth of sproutsin shelterswas again ap- The seedlingsproutsfrom the medium size classgrowing in shelters had more than double the height growthof theircounterparts outsideof shelters in the first growing season (Fig.1 Table1). The meanheightat the end of the first growing seasonobscures some of the more dramatic re- suits. The maximum height growth observedfor a seedling sprout in a shelter was 183.0 cm, whereas the maximumheight for a sproutoutside of a shelterwas 51.0 cm. Althoughtotal heightsinsideand outsideof shelterswere significantlydifferentat the end of the secondgrowingseason,the height growth that year was not significantly different. Nine of the 32 shelteredsproutsgrew lessthan 10cm in height. The seedlingsthat experienced poor height growth had been observed to exhibit second and third flushes of height growth during the end of the previousgrowing season. Such active height growth may have been inadequatelyhardened off and was possiblydamagedby frost in the fall. Basal diameters were again consistentlygreater for sprouts growing riod. proximatelydouble that of sprouts outside of sheltersduring the first growing season (Table 1). Height growth during the secondyear was much lower for all sprouts,and there was no significant difference in growthinsidecomparedto outsideof shelters.Some sproutshad emerged from the topsof sheltersby the end of the first growing season, and the mean height was above the 150 cm heightof the sheltersat the endof the secondgrowingseason.Basaldiameters were significantly greater for sproutsoutsideof sheltersat the end of the secondgrowing season. ShootGrowth Patternsof Seedlings and Sprouts The total height growth of both seedlings(unclippedtreatment) and sprouts (clipped treatment) inside shelterswas significantlygreaterthan that outside of shelters (Table 2). Shoot elongationpatternsvaried between sproutsand seedlings(Fig. 2). Many treesin all four treatmentsunderwenttwo flushesof shootelonga- tion during the growingseason,with most growth occurringin the initial flush. Outsideof shelters,sproutshad greaterheightgrowthanda longerduration of shootelongationduring the initial flush, comparedto seedlings. The samepattern existedfor treesinside shelters, but with both the dura- tion and magnitude of extension growth being much greaterthan outside shelters(Table 2). With sprouts,the number of leaves producedon the leaderduring the initial flush was similar inside and out- sideof shelters.With seedlings,leader shoots inside shelters had greater numbers of leaves; this was due to two trees inside of shelters developing large numbersof leaves(24 and 35). The other trees inside shelters had a mean of 11.5 leaves, which was similar to those not in shelters. Seedlings(i.e., nonclippedindividuals)essentiallyhad a 73-cmheadstart in total height (seeTable 2). As a result,totalheightaccumulated by seed- lingsand sproutsinsideshelterswas similarat the end of the first growing season,in spite of the greater height growth of the sprouts.At the end of two growingseasons,seedlingsinside sheltershad significantlylarger basal diametersthan sproutsinside shelters [12.9mm (s.e. = 0.4) for seedlings,9.9 mm (s.e. = 0.7) for sprouts;t-test:P < 0.011. CompetingVegetationand Mortality Prior to establishingthe deer exclosuresat both sites,understoryvegeta- tion was preventedfrom developing aboveabout 100cm in heightby deer browsing. When protected from browsing,regenerationdominatedby red maple and several birch species quicklybecameestablishedand grew rapidlyin height.The oak sproutsand seedlingsin sheltersin the medium and large size classesgenerallymaintained dominant or codominant positions in the partially dosed seedling canopythat developedin the first 3 years. Sproutmortalitywas highestin the medium sizeclass,and was higherfor sproutsin sheltersthan thoseoutside. Some was caused by damage to sprouts when shelters were blown down by wind; othercausesappeared to be related to frost damage,insect damage,and additionalunknownfactors. DISCUSSION The greatestheight growth ratesoccurred Fig. 1. An exampleof a red oak sproutfrom the medium size classat the end of the first growing season.The tree had been clipped down to ground level the previousdormant season,the shelter was installed immediately after clipping. in established natural red oak regenerationwith basaldiametersof 8 to 15 mm (i.e., medium and large size classes)that had been clippedback to the groundand protectedby tree shelters. Sprouts in sheltershad double NJAF9(1992) 143 shelters 2OO 180-rS'•ROUT/SHELTœR determtnate berof leaves,butinternodes elongated a greateramountovera longerperiod of time. Only two seedlingsclearly shiftedto indeterminategrowth, producingmanylate leaves.Sproutsoutsideof sheltersgrew for a meanof 36 daysin the initialflush,producing17 leaves. In shelters, the period of growthfor sproutswasextendedto 52 days,with the samenumberof leaves, the greaterheight growth was agatn the resultof greaterinternodeelonga- •140•E•LIN•••,• oE•120T matntatned growthand producedthe samenum- 100- tion. Many seedlingsand sprouts,both 40- • 20 ..... • • • .... inside and outside of shelters, formed '............. terminal buds midway through the growingseasonand then produceda secondflushof growthlaterin the season.The secondflusheswere longer •----•----•'"'•""• for trees inside of shelters, whtch 0 6 12 19 26 33 40 GROWING 47 54 SEASON 62 68 75 82 89 96 addedto the totalheightdifferences at the end of the growingseason(Table DAYS 2). However, the differencesin total FiS.2. Cum•a•e heishtS•o•h of te•mia• shoots ]•e size •s duda• •e fi•st •owia• season.Sproutshoots•ew •mm buds on met co]]• o• stems•at had beendipped backto Smuad]eve]•e p•eviousdemur season; seed]ia• shoo• •ew ffo• •e te•Jaa] buds of stemsthat had not been dipped. •easuremeatsbe• on •9 •ay (•m•a• seasonday 0), and extendedua•] 23 Au•st (•owJaSsea•a day 96). Bo• seed]ia• had •o flusheswi• dete•Jaate •ow• flush. •e sproutwith no she]te•had one flush wJ• Jadete•iaate Sm•h ]asQaSuatJ] day33. •e sproutiasidea she]te•had one•ush o• iadetemiaate •mw• ]astia• ua•] day •, followed by a secondflush •m• day 82 to 89. flush of trees inside shelters. because of differences leaves contained expand.However, sproutsand rapidly growingseedlings mayexhibitindeterminategrowth, in which an in- ameters were smaller than sprouts as the shootexpands;the numberdependsupon growingconditionsand the amountof storedcarbohydrate reservein the rootsystem(Wilson1984). In thepresentexperiment,seedlings growingoutsideshelters.Thesetall, small-diameter stems were unable to in the basic growth patternbetweensproutsand seedlingsoutsideof shelters(sprouts initially have indeterminategrowth and producemore leavesand intern- shelters,and somegrew beyondthe top of the 150-cmsheltersduring the firstgrowingseason. Theirstemswere straightbutflexible,andtheirbasaldi- number The ef- fect on sproutsappearsto be greater the height growth of those outside definite in the terminal heightgrowthin the firstgrowingseason were largely due to increasesin the durationand magnitudeof internode elongation during the initial bud odes than do the initial flushes of seedlings). The two basicdifferences in growth of leaves are formed of trees inside shelters measured tn the present study (i.e., increased heightgrowthand decreased basaldtametergrowth)havebeenobserved•n sessileoakand pedunculateoak (Rendie 1985,Tuley1985).It mayseemlog- supportthemselves and requiredthe lateral support of the sheltersto reoutside of shelters had an initial mainupright. For sproutsfrom smallerseedlings growth flush that lasteda mean of (initial basaldiameterof 5 mm), shelonly 17 days, with 11 leaveson the ters did not significantlyimprove leader. Most of the seedlingsinside height growth over 3 growing sea- ical to conclude that the taller stems of oaksin sheltersindicatea greatertotal dry weight production,but Rendle sons, but diameters were still consis- tently smallerfor sproutsinsideshelters. The lack of improved height growth in thesesproutsmay have resuitedfromtheir smallerroot systems. Studieshaveshownthatoakseedlings Table2. Mean measures of the initialgrowthflush(duration,height,and numberof leaves)andtotal growthduringthe firstgrowingseasonfor sproutsandseedlings of the largesizeclass(initialbasaldiameter= 15 mm).a Sprouts allocate most of their resources in the Shelter firstfew yearsto the development of a vigorousroot system;only when an n individual Duration of has achieved substantial root developmentwill it respondto overstoryremovaland growin height (Sanderet al. 1984).Our resultssuggestthat sheltersdonot alterthisbasic growth pattern. Inherent differencesin growth patternsbetweenseedlingsand sprouts influenced the level of responseto shelters.Oaksgenerallyhavedeterminateshootgrowth,with a shortperiod of extensiongrowth in which the internodeselongateand the preformed 144 NJAF9(1992) Initial ht (cm) initialflush (days) Heightof initial flush (cm) No. of leaves on initial flush Seedlings No shelter 12 12 0 0 51.8a (2.7) 113.0a Shelter No shelter 12 12 72.8 73.3 36.1b 32.2b 17.0c (2.5) 60.8b (5.3) 52.0b (0.8) 26.2c (9.4) (5.1) (8.3) (3.5) 19.8a (2.1) 17.3a,b (1.6) 14.5b,c (2.2) 10.9c (0.9) No. of trees with second flush Total ht growth 9 136.8a 4 67.4b 9 73.8b 7 38.2c for season (cm) (11.0) (6.0) (8.3) (3.8) Final ht at end 136.8a,b 67.4c 146.6a 111.5b of season (cm) (11.0) (6.0) (8.3) (3.8) Meansin each row followedby differentlettersare significantly different(P < 0.05), usingTukey's studentizedrangetest. Standarderrorsare listedin parenthesesbeneatheach mean. hardwoodstandswherelargered oak naturalregenerat,onex,sts,but ,s prevented from developingin height by that differences existed in allocation of browsingherbivores.Clippinglargeproduction. diameterseedlings(minimumbasal diametersof about8 mm) at ground The growth changesobservedin shelters aregenerallyascribed to alter- level during the dormantseasonand protecting the resulting seedling ataonsin temperature,humidity, and carbon dioxide levels. However, it is sproutswith shelterscausedoaks to alsopossiblethat the protectionfrom growabovethe rangeof deerbrowse in 1 to 3 years. The managementcost w,ndprovidedby sheltersmayalsoaffectgrowthallocation.In greenhouse of this applicationof sheltersis less than for most uses, since there is no experiments with sweetgum(Neeland Harris1971,1972),seedlings thatwere investment in planting stock, site shaken to simulate wind effects had preparation, or planting. It may be greaterheight growth,less diameter possibleto ensurethe dominanceof growth, and longerperiod of exten- an oak componentin these stands sion growth prior to settingterminal with relatively few shelteredseedbuds,comparedto treesthat remained lings, perhapsas few as 100-200/ha. motionless. Similar differences were Redoakformsa nearlypureoverstory seen between staked and unstaked canopywith as few as 100trees/hain field-grownseedlings.The physiolog- mature stands (Oliver 1978), and has ,cal mechanismof this responsehas relativelylow mortalityratesfrom the notbeenestablished, but it mayresult sapling stage onward. Estimatesof fromincreasedproductionof ethylene mortalityratesfor the transitionfrom m swaying stems (Kramer and Kozthe shelteredsproutstageto the saplowski1979). ling stagewouldbe neededin orderto Sheltersin thisstudywerebeneficial developguidelines for minimumnumbers of shelters needed. These estito height growth in the medium and mates would best be obtained from trilargesizesonly duringthe first growalsof sheltersusedwith widelyspaced ing season.In the secondyear, when oaks,encompassing the rangeof varimanyof the stemshad emergedfrom the tops of the shelters,the average ationsnormallyfound under operational conditions, rather than from growthwas the sameas for treesoutside shelters.At that point, microcli- studiesof spatiallygroupedregeneration of uniformsizes,asreportedhere. maticeffectswould no longerinfluTwo variations on this approach ence the developingterminal shoot; effectsof wind on shakingthe stem alsohave potential.First, smalleroak seedlings canbeclippedandprotected wouldalsobeginto occur.Heightand d,ametergrowthratesand biomassalby shelters.Althoughheightgrowth locationpatternsreturn to levelsthat responsesdid not occurin the first 3 approach"normal"(i.e., similarto ungrowingseasonsfor sproutsfrom the shelteredtrees)within about 2 years smallersizeclassin the presentstudy, (Tuley 1985, Potter 1991). It is clear they were effectivelyprotectedfrom that sheltersmustbe left in placefor browsingand maygraduallyreachthe support at least 2 years after trees sizewhereheightgrowthis markedly emergefrom the topsof shelters,until acceleratedby the shelter environthey developsturdystems. ment. Second,sheltersmaybe placed around the seedlings themselves, APPLICATION rather than the cut stump. Height growth will likely be less than for havea more The resultsof this study indicate sprouts,but the seedhngs stablestem.An importantfactorin dethat a potentialexistsfor the use of tree sheltersin New Englandmixed- cidingbetweenseedlingsand sprouts foundthat total dry weightsof seed1,ngs,ns,de and outs,de of shelters were not significantlydifferent, but would be how badly deformedthe ex- ,stangseedling,s, as a resultof repeatedbrowsing. LITERATURE [] CITED CARMEA•/, W.H. 1978. Site index curves for northern hardwoods in northern Wisconsin and upper Michigan. USDA For. Serv. Res. Pap.NC-160.16 p. FRUARSO•, K., A•D N.D. WEISS.1987.Improved growthrateswithintreeshelters.Quart.J. For. 81(3):184-187. HORSLEY, S.B. 1988. How vegetationcan influenceregeneration. p. 38-55.in Proc.Guidelines for regenerating Appalachian hardwood stands. SAF Pub. No. 88-03. Soc. Am. For, Bethesda, MI. KrrTR•DGE,D.B., JR., A•D P.M.S. AShTOn/.1991. White-taileddeer (Odocoileus virginianus) and theirbrowseeffecton seedlingregeneration of mixed speciesstandsin southernNew England.In prep. I•a•MSR,P.J.,A•D T.T. Kozrowsra.1979.Physiologyof woodyplants.AcademicPress,New York. 811 p. LANTAGNE,D.O., C.W. RAMM, AND D.I. DICI•- MA•. 1990.Tree sheltersincreaseheightsof plantedoaksin a Michiganclearcut.North. J. App1.For. 7(1):24-26. NEar, P.L., A•D R.W. HARRIS. 1971. Motion- inducedinhibitionof elongationand induction of dormancyin Liquidambar. Science173:58-59. NEEt, P.L., A•D R.W. HARRIS.1972. Tree seed- ling growth:Effectsof shaking.Science175: 918-919. Or•VER,C.D. 1978.The development of northern red oak in mixed stands in central New En- gland.YaleUniv. Sch.For. Bull.No. 91. 63 p. PooreR,M.J. 1988.Treeshelters improvesurvival and increaseearly growthrates.J. For. 86(8): 39-41. PoTz•, M.J. 1991. Treeshelters. For. Comm. Handb. No. 7. 48 p. R•OrE, E.L. 1985. The influence of treeshelters on microclimate and the growthof oak. In Proc. 6thmeetingNationalHardwoods Prog.Oxford For. Inst., Oxford, UK. SA•D•R, I.L., P.S. Joa•so•, A•D R. RODGERS. 1984.Evaluatingoak advancereproductionin the Missouri Ozarks. USDA For. Serv. Res. Pap.NC-251.16 p. TULLY,G. 1983.Sheltersimprovethe growthof youngtreesin the forest.Quart. J. For. 77(2): 77-87. TULLY,G. 1985.The growthof youngoaktreesin shelters.For. 58(2):181-195. USDA SO•L CONSERVATION SERVICE. 1967. Soil Survey,FranklinCounty,Mass. USDA So•[ CO•SERVAT•O• SERWC•. 1981. Soil Surveyof WindhamCounty,Conn.130p. WILSO•, B.F. 1984. The growing tree. Rev. ed. Univ. of Mass.Press,Amherst.138p. W•DULL, K.N. 1991.Tree sheltersfor seedling protection.USDA For. Serv. Tech.and Dev. Prog.Rep.Seedlingprotection. 142p. NJAF9(1992) 145